Re: [PATCH v4 5/5] x86: KVM: SEV: implement kvm_lock_all_vcpus and use it

From: Peter Zijlstra <peterz@infradead.org>
To: Sean Christopherson <seanjc@google.com>
Cc: Maxim Levitsky <mlevitsk@redhat.com>,
	kvm@vger.kernel.org, linux-riscv@lists.infradead.org,
	Kunkun Jiang <jiangkunkun@huawei.com>,
	Waiman Long <longman@redhat.com>,
	linux-kernel@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	Catalin Marinas <catalin.marinas@arm.com>,
	Bjorn Helgaas <bhelgaas@google.com>,
	Boqun Feng <boqun.feng@gmail.com>, Borislav Petkov <bp@alien8.de>,
	Albert Ou <aou@eecs.berkeley.edu>,
	Anup Patel <anup@brainfault.org>,
	Paul Walmsley <paul.walmsley@sifive.com>,
	Suzuki K Poulose <suzuki.poulose@arm.com>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	Alexandre Ghiti <alex@ghiti.fr>,
	Alexander Potapenko <glider@google.com>,
	Oliver Upton <oliver.upton@linux.dev>,
	Andre Przywara <andre.przywara@arm.com>,
	x86@kernel.org, Joey Gouly <joey.gouly@arm.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	kvm-riscv@lists.infradead.org,
	Atish Patra <atishp@atishpatra.org>,
	Ingo Molnar <mingo@redhat.com>,
	Jing Zhang <jingzhangos@google.com>,
	"H. Peter Anvin" <hpa@zytor.com>,
	Dave Hansen <dave.hansen@linux.intel.com>,
	kvmarm@lists.linux.dev, Will Deacon <will@kernel.org>,
	Keisuke Nishimura <keisuke.nishimura@inria.fr>,
	Sebastian Ott <sebott@redhat.com>,
	Shusen Li <lishusen2@huawei.com>,
	Paolo Bonzini <pbonzini@redhat.com>,
	Randy Dunlap <rdunlap@infradead.org>,
	Marc Zyngier <maz@kernel.org>, Zenghui Yu <yuzenghui@huawei.com>
Subject: Re: [PATCH v4 5/5] x86: KVM: SEV: implement kvm_lock_all_vcpus and use it
Date: Sat, 3 May 2025 12:08:20 +0200	[thread overview]
Message-ID: <20250503100820.GF4198@noisy.programming.kicks-ass.net> (raw)
In-Reply-To: <aBUxqZ6qgfYZLsye@google.com>

On Fri, May 02, 2025 at 01:57:13PM -0700, Sean Christopherson wrote:

> int kvm_lock_all_vcpus(struct kvm *kvm)
> {
> 	struct kvm_vcpu *vcpu;
> 	unsigned long i, j;
> 	int r;
> 
> 	lockdep_assert_held(&kvm->lock);

So I agree that having this assertion here is probably good from a
read-code pov, however, strictly speaking, it is redundant in that:

> 	kvm_for_each_vcpu(i, vcpu, kvm) {
> 		r = mutex_lock_killable_nest_lock(&vcpu->mutex, &kvm->lock);

will implicitly assert kvm->lock is held. If you try to use an unheld
lock as nest lock, it will complain loudly :-)

(my inner pendant had to reply, ignore at will :-)

> 		if (r)
> 			goto out_unlock;
> 	}
> 	return 0;
> 
> out_unlock:
> 	kvm_for_each_vcpu(j, vcpu, kvm) {
> 		if (i == j)
> 			break;
> 		mutex_unlock(&vcpu->mutex);
> 	}
> 	return r;
> }
> EXPORT_SYMBOL_GPL(kvm_lock_all_vcpus);

-- 
kvm-riscv mailing list
kvm-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kvm-riscv